Oddball Ping Results

[rmmagow]

Solaris 8 on an ES450.
The box crashed due to a disk problem. It rebooted itself to the "run fsck manually" level i.e. waiting for the root password to perform system maintenance. When the machine crashed I was called up and attempted to see what was going on. I used my VPN connection and attempted to ping the device and got a failure, no surprise here. I then telnetted to the router that serves as the gateway for the failed machine and pinged from there. Much to my surprise the command returned normally. Telnet from the router to the failed machine didn't work (no surprise). Question, why'd it reply to ping and did it reply to ping or did the router (which still had an ARP entry for the box) simply figure all was OK and reply on behalf of the dead box. If this is the case, is this how the RFC has ping working on a local subnet or should I contact my router vendor and open a ticket on something like this. More curious than anything. Thanks.

 

[jaytco]

While I don't have the tech specs. I have found that ping will work while the interface is up, but not necessarly all the software of the box. As far as the ping from the local subnet and outside the subnet. With the ping on the local subnet, the interface does not need a default route, while outside the subnet it will. --Jay

 

[rmmagow]

And thank you. Makes perfect sense in hind-sight. I did see the adaptor come up then the machine stopped on the drive problem before I saw the messages regarding the gateway.
Thanks for such a quick and through answer.

 

[Arluin]

The router probably pings the adapter by Layer-3 protocol that the adapter can handle while the other ping attempts used higher protocols that required the OS to respond to.